Component and Total Forest Aboveground Biomass Estimation Using GF-1/3 Images

Accurately estimating forest component above-ground biomass is crucial for understanding and analyzing the growth process of vegetation ecosystems, interpreting the ecosystem carbon cycle correctly, and improving the effectiveness of forest management. It also provides ideas for the bottleneck problem of low saturation point in forest aboveground biomass (AGB) estimation using remote sensing techniques. This paper extracts a large number of remote sensing features from optical GF-1 and GF-3 SAR data. A feature optimization inversion model named KNN-FIFS (a fast iterative procedure embedded in K-nearest neighbor) was applied here for component and total forest AGB inversion. Four forest types including Yunnan pine, Simao pine, broadleaf forest, and mixed conifer and broadleaf forest were involved in this study. The results indicate that the summation of the estimated each component AGB for total forest AGB estimation showed higher inversion accuracy than the total AGB estimated directly by remote sensing features. The relative RMSE difference between them was about 3%. The results also revealed that through forest component estimation for later total forest AGB estimation can help improve the saturation points in forest AGB estimation using remote sensing datasets.